https://gcc.gnu.org/bugzilla/show_bug.cgi?id=118570
Richard Biener <rguenth at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Ever confirmed|0 |1
Status|UNCONFIRMED |NEW
CC| |hubicka at gcc dot gnu.org,
| |rguenth at gcc dot gnu.org
Component|tree-optimization |ipa
Last reconfirmed| |2025-01-21
--- Comment #3 from Richard Biener <rguenth at gcc dot gnu.org> ---
I think we have a duplicate somewhere about const/pure functions vs. (early)
inlining, also with respect to CSE. This testcase shows it might be also
important for late inlining because of invariant motion. With mod-ref it
also extends to const/pure with other (known but irrelevant) side-effects.
So confirmed, though indeed this benchmark outer loop isn't practically
relevant to the extent of "cheating".
